Gradient Based Activations for Accurate Bias-Free Learning

TL;DR

This paper introduces a gradient-based feature masking technique that leverages biased discriminators to effectively reduce bias and improve accuracy in machine learning models, outperforming existing adversarial methods.

Contribution

It proposes a novel gradient-based feature masking approach that uses biased discriminators to enhance bias mitigation without sacrificing accuracy.

Findings

Reduces bias effectively while improving accuracy.

Outperforms recent bias mitigation methods on benchmarks.

Enhances unbiased feature learning through gradient-based masking.

Abstract

Bias mitigation in machine learning models is imperative, yet challenging. While several approaches have been proposed, one view towards mitigating bias is through adversarial learning. A discriminator is used to identify the bias attributes such as gender, age or race in question. This discriminator is used adversarially to ensure that it cannot distinguish the bias attributes. The main drawback in such a model is that it directly introduces a trade-off with accuracy as the features that the discriminator deems to be sensitive for discrimination of bias could be correlated with classification. In this work we solve the problem. We show that a biased discriminator can actually be used to improve this bias-accuracy tradeoff. Specifically, this is achieved by using a feature masking approach using the discriminator's gradients. We ensure that the features favoured for the bias discrimination are de-emphasized and the unbiased features are enhanced during classification. We show that this simple approach works well to reduce bias as well as improve accuracy significantly. We evaluate the proposed model on standard benchmarks. We improve the accuracy of the adversarial methods while maintaining or even improving the unbiasness and also outperform several other recent methods.